Investigating the rate of photoreductive glucosyl radical generation

Article abstract of DOI:10.1021/ol200644w

Chemical equations presented. The photoreduction of glucosyl halides to generate glucosyl radicals has been investigated to probe the nature of the photoredox cycle. Amine (the reductant) and catalyst concentration affect the reaction rate at low concentrations but exhibit saturation at higher concentrations. Water and hydrophobic catalysts were found to significantly increase the conversion efficiency.

Full text of DOI:10.1021/ol200644w

ORGANIC
LETTERS
2
011
Vol. 13, No. 9
406–2409
Investigating the Rate of Photoreductive
Glucosyl Radical Generation
2
†
‡
,†
R. Stephen Andrews, Jennifer J. Becker, and Michel R. Gagn ꢀe *
University of North Carolina at Chapel Hill, Department of Chemistry, Chapel Hill,
North Carolina 27599, United States, and U.S. Army Research Office, P.O. Box 12211,
Research Triangle Park, North Carolina 27709, United States
mgagne@unc.edu
Received March 10, 2011
ABSTRACT
The photoreduction of glucosyl halides to generate glucosyl radicals has been investigated to probe the nature of the photoredox cycle. Amine
the reductant) and catalyst concentration affect the reaction rate at low concentrations but exhibit saturation at higher concentrations. Water and
hydrophobic catalysts were found to significantly increase the conversion efficiency.
(
Transition metal photocatalysts have generated signifi-
cant interest as powerful single electron transfer (SET)
1
reagents that operate under mild conditions. Previously
recently in a variety of synthetic reactions, including
3
4
reductive dehalogenation, [2 þ 2] cycloadditions, and
5
radical additions into unsaturated bonds. The most pro-
developed as a catalyst for the capture and conversion
2þ
mising aspect of this chemistry is the mild reaction condi-
tions, generally requiring only a stoichiometric redox
reagent and irradiation from household fluorescent bulbs,
LEDs, or sunlight. Thus, photoredox catalysts have the
exciting potential to convert sunlight, an abundant and
cheap light source, into chemical energy for driving useful
synthetic transformations.
2
of solar energy, Ru(bpy)
has been employed more
3
†
University of North Carolina at Chapel Hill.
U.S. Army Research Office.
‡
(
1) For recent reviews, see: (a) Zeitler, K. Angew. Chem., Int. Ed.
009, 48, 9785–9789. (b) Narayanam, J. M. R.; Stephenson, C. R. J.
Chem. Soc. Rev. 2011, 40, 102–113.
2) (a) Kalyanasundaram, K. Coord. Chem. Rev. 1982, 46, 159–244.
b) Juris, A.; Balzani, V.; Barigelletti, F.; Campagna, S.; Belser, P.; von
Zelewsky, A. Coord. Chem. Rev. 1988, 84, 85–277.
3) (a) Narayanam, J. M. R.; Tucker, J. W.; Stephenson, C. R. J.
2
(
The photophysical properties and reaction chemistry of
2
(
2þ
Ru(bpy)3 has been extensively studied. Crudely speak-
ing, absorption into a metal to ligand charge transfer
(
J. Am. Chem. Soc. 2009, 131, 8756–8757. (b) Maji, T.; Karmakar, A.;
Reiser, O. J. Org. Chem. 2011, 76, 736–739.
(
MLCT) band (λmax = 452 nm) generates an excited state
2þ
complex (*Ru(bpy)3 ) that can be chemically quenched
to form the strong oxidant Ru(bpy)
(
4) (a) Ischay, M. A.; Anzovino, M. E.; Du, J.; Yoon, T. P. J. Am.
Chem. Soc. 2008, 130, 12886. (b) Du, J.; Yoon, T. P. J. Am. Chem. Soc.
009, 131, 14604. (c) Ischay, M. A.; Lu, Z.; Yoon, T. P. J. Am. Chem.
Soc. 2010, 132, 8572–8574.
5) (a) Nicewicz, D. A.; MacMillan, D. W. C. Science 2008, 322, 77–
0. (b) Andrews, R. S.; Becker, J. J.; Gagn ꢀe , M. R. Angew. Chem., Int.
3
3
þ
or the strong
2
þ
reductant Ru(bpy)3 . Despite the potential for providing
unexplored venues for improving existing methods and
discovering new ones, the applicability of these photo-
physical studies to synthetic protocols has not been
established.
We recently reported the intermolecular coupling of
electron deficient alkenes and glucose C1-radicals, the
latter being generated by the reduction of R-glucosyl
(
8
Ed. 2010, 49, 7274–7276. (c) Condie, A. G.; Gonz ꢀa lez-G oꢀ mez, J. C.;
Stephenson, C. R. J. J. Am. Chem. Soc. 2009, 132, 1464–1465.
(d) Tucker, J. W.; Narayanam, J. M. R.; Krabbe, S. W.; Stephenson,
C. R. J. Org. Lett. 2009, 12, 368–371. (e) Furst, L.; Matsuura, B. S.;
Narayanam, J. M. R.; Tucker, J. W.; Stephenson, C. R. J. Org. Lett.
2
010, 12, 3104–3107. (f) Tucker, J. W.; Nguyen, J. D.; Narayanam,
J. M. R.; Krabbe, S. W.; Stephenson, C. R. J. Chem. Commun. 2010, 46,
985–4987. (g) Nagib, D. A.; Scott, M. E.; MacMillan, D. W. C. J. Am.
4
þ
Chem. Soc. 2009, 131, 10875. (h) Shih, H.-W.; Vander Wal, M. N.;
Grange, R. L.; MacMillan, D. W. C. J. Am. Chem. Soc. 2010, 132,
bromide by photogenerated Ru(bpy)3 . While high yield-
ing, this protocol was characterized by long irradia-
13600–13603. (i) Nguyen, J. D.; Tucker, J. W.; Konieczynska, M. D.;
5
b
Stephenson, C. R. J. J. Am. Chem. Soc. 2011, 133, 4160–4163.
tion times (18ꢀ48 h). Since numerous researchers have
1
0.1021/ol200644w r 2011 American Chemical Society
Published on Web 04/07/2011

demonstrated that photoexcitation and quenching of Ru-
(
electron transfer (ET) processes, we considered the
possibility that the source of our long reaction times was
a nonproductive conversion of light into heat.
reaction conditions affected the steady state concentration
þ
2þ
bpy)3 is reversible and subject to energy-wasting back
of Ru(bpy)3
.
2
,6
Scheme 1. Photoreduction of Glucosyl Radical
Figure 1. Amount of 1 consumed after 3 h of irradiation with a
4 W CFL source versus initial concentration of 1. The dashed
vertical line represents the approximate concentration for our
1
5
b
radical addition reactions. Reaction conditions: 6.0 mM Ru-
(
2þ
i
, 0.37 M EtN Pr , 0.24 M BuSH in CD CN.
t
bpy)
3
2
3
Varying the catalyst concentration revealed that un-
der dilute conditions (<4 mM) the rate of the reaction
In the reaction mechanism in Scheme 1, we presumed
that the reaction inefficiencies were localized in the photo-
redox cycle, as C1-radical trapping by methyl acrylate was
expected to be fast and efficient. We therefore set out to
probe the mechanism of the photoredox cycle and either
support or refute our hypothesis. To simplify the overall
2
þ
was proportional to [Ru(bpy)3 ] (Figure 2), though at
concentrations >4 mM the rate plateaued at a flux
dependent efficiency. We rationalized this saturation
to result from a photon limited scenario wherein the
catalyst concentration was higher than the photon flux.
transformation we utilized a reductive trap for the radical
(
t
i.e., BuSH) in lieu of the alkene. Conditions similar to our
previously described reactions (R-glucosyl bromide (1),
Ru(bpy)3 , and N,N-diisopropylethylamine as stoichio-
metric reductant) were found to readily debrominate
2þ
3
glucosyl bromide to 2 (Scheme 1). Since these simple
a
reactants and products were amenable to direct quantita-
7
1
tion by H NMR spectroscopy, the conversion efficiency
of test reactions could be readily obtained from irradia-
8
tions in NMR tubes. This procedure provided the means
to readily compare reaction rates as a function of experi-
mental conditions.
Importantly, control experiments indicated that reac-
8
tion conversions were independent of thiol concentration,
indicating that radical trapping was fast and that the
conversions reflected the rate at which glucosyl radical
was being generated. In contrast, the rate was dependent
on [1]ini (Figure 1), suggesting a turnover limiting SET
Figure 2. Conversion after 3 h of irradiation with increasing
[Ru(bpy)3 ] for CFL and blue LED light sources. The dashed
vertical line represents the catalyst concentration for our radical
2þ
þ
5b
i
from the reduced catalyst Ru(bpy)3 to 1 (step a, Scheme 1).
Since the rate equation for this step would be dependent
2
addition reactions. Reaction conditions: 0.37 M EtN Pr ,
t
0
3
.24 M BuSH, 0.12 M 1 in CD CN.
þ
on both the concentration of 1 and Ru(bpy)3 , our atten-
tion shifted toward designing experiments to probe how
Consistent with this interpretation were the higher con-
versions of reactions using blue LED strips (λemission
35 ( 15 nm), which have been reported to create
a higher flux at the MLCT absorption wavelength
=
4
(
6) (a) Tazuke, S.; Kitamura, N.; Kawanishi, Y. J. Photochem. 1985,
9, 123–138. (b) Bock, C. R.; Connor, J. A.; Gutierrez, A. R.; Meyer,
T. J.; Whitten, D. G.; Sullivan, B. P.; Nagle, J. K. Chem. Phys. Lett.
979, 61, 522–525. (c) Monserrat, K.; Foreman, T. K.; Graetzel, M.;
Whitten, D. G. J. Am. Chem. Soc. 1981, 103, 6667–6672.
7) Dispersion in the chemical shifts of 1 and 2 enabled independent
2
9
þ
Figure 2). Since the Ru(bpy) generation also requires
3
(
1
a stoichiometric reductant to intercept the excited state,
i
we varied the concentration of EtN Pr and similarly
(
2
quantitation through the H5 protons on the glucosides by comparison
with 1,3,5-trimethoxybenzene as an internal standard.
found that the reaction exhibited saturation behavior
(Figure 3).
(
8) See Supporting Information for control experiments.
Org. Lett., Vol. 13, No. 9, 2011
2407

highly sensitive to solvent effects, especially water, this
appeared to be a logical place to look for increased
5
i,12
efficiencies.
Figure 3. Conversions after 3 h of irradiation as a function of
i
initial EtN Pr
the amine concentration for our radical addition reactions.
2
concentration. The dashed vertical line represents
5
b
2þ t
, 0.24 M BuSH, 0.12 M
Reaction conditions: 5.9 mM Ru(bpy)
in CD CN.
3
1
3
Figure 4. % conversion of 1 (20 min) after photolysis with blue
,
2
3
þ
LEDs versus cosolvent additive. Conditions: 5.9 mM Ru(bpy)
0.37 M EtN Pr , 0.24 M BuSH, and 0.12 M 1 in CD CN.
2 3
i
t
The observation that both catalyst and amine show
saturation behavior near the experimental conditions sug-
gested finely balanced rates for the elementary steps lead-
To determine if solvation effects were limiting our
photoredox cycle, we investigated the effect of water,
ethylene glycol, methanol, and DMF as cosolvents in
acetonitrile. As shown in Figure 4, water dramatically
enhanced this conversion efficiency (note that the time
points are 20 min in Figure 4 compared to 3 h in
Figures 1ꢀ3). Methanol and DMF (not shown) were
ineffective while ethylene glycol was modestly effective.
Unexpectedly, the conversion dependence on catalyst
concentration was markely different from anhydrous con-
þ
ing to Ru(bpy)3 . In the search for approaches to improve
þ
the overall efficiency of Ru(bpy)3 generation, the bimo-
2þ
lecular reductive quenching of *Ru(bpy)3 was considered
to be the most chemically manipulable as it has been
previously described in typically Marcus terms, proceeding
via an encounter complex (van der Waal) that precedes ET
6
,10
and successor complex resolvation (Scheme 2).
2þ
ditions as [Ru(bpy)3 ] < 1 mM led to strong linear
increases with catalyst concentration, while this behavior
inverted beyond this threshold (Figure 5).
Scheme 2. Resolvation of Catalyst
Since the efficiency of the overall quenching is related to
the competing rates of back ET (k ) and ion solvation
bt
6
,10,11
(
k ),
esc
and the latter process has been shown to be
(9) LED strips purchased from www.creativelightings.com. See also
ref 5e.
(
10) (a) Kitamura, N.; Kim, H. B.; Okano, S.; Tazuke, S. J. Phys.
Figure 5. Conversion after 20 min of irradiation with blue LEDs
Chem. 1989, 93, 5750–5756. (b) Ballardini, R.; Varani, G.; Indelli, M. T.;
Scandola, F.; Balzani, V. J. Am. Chem. Soc. 1978, 100, 7219–7223.
(
2þ
i
with increasing [Ru(bpy)
0.22 M BuSH, 0.11 M 1 in 10:1 CD CN/H O.
3
]. Reaction conditions: 0.33 M EtN Pr
2
,
t
11) For medium effects in electron transfer processes, see: (a) Sun,
3
2
H.; Yoshimura, A.; Hoffman, M. Z. J. Phys. Chem. 1994, 98, 5058–5064.
(
(
(
1
4
b) Clark, C. D.; Hoffman, M. Z. J. Phys. Chem. 1996, 100, 7526–7532.
c) Clark, C. D.; Hoffman, M. Z. J. Phys. Chem. 1996, 100, 14688–14693.
d) Fedurco, M.; Sartoretti, C. J.; Augustynski, J. J. Phys. Chem. B 2001,
05, 2003–2009. (e) Kavarnos, G. J.; Turro, N. J. Chem. Rev. 1986, 86,
01–449. (f) Julliard, M.; Chanon, M. Chem. Rev. 1983, 83, 425–506.
Using a similar set of principles as Scheme 2, it has been
shown that more hydrophobic photocatalysts will resol-
vate more quickly after reductive quenching, which re-
(
g) Kitamura, N.; Kim, H. B.; Okano, S.; Tazuke, S. J. Phys. Chem.
989, 93, 5750–5756. (h) Ballardini, R.; Varani, G.; Indelli, M. T.;
Scandola, F.; Balzani, V. J. Am. Chem. Soc. 1978, 100, 7219–7223.
13
duces the competitiveness of back-electron transfer (k ).
bt
1
2þ
0
To test this, Ru(dmb)3 (dmb = 4,4 -dimethyl bpy) was
2
408
Org. Lett., Vol. 13, No. 9, 2011

Scheme 3. Radical Mediated CꢀC Bond Formation
Figure 6. Conversion after 20 min of irraditation with blue LEDs
for Ru(L)
EtN Pr , 0.24 M BuSH, 0.12 M 1 in CD CN/H O.
2 3 2
2
þ
2þ
3
. Reaction conditions: 6.0 mM Ru(L) , 0.37 M
3
In summary, we have demonstrated that the rate of the
photoredox cycle is dependent on several variables, includ-
ing the hydrophobicity of the catalyst and the presence of
polar cosolvents. Applying these observations to a known
reaction demonstrated the applicability of these mechan-
istic studies to improving problems of synthetic relevance.
Current work focuses on applying these concepts to a
wider variety of photoredox reactions.
i
t
used as the catalyst at varying concentrations of water. As
shown in Figure 6 a significantly more effective transfor-
mation was observed that required only 2% water to achieve
plateau behavior.
This more efficient catalyst was applied to the radical
mediated coupling of 4 and methyl acrylate. As hoped for,
2þ
a rate acceleration was observed for Ru(dmb)3 , which
provided full conversion of 4 under anhydrous conditions
in 6 h. In contrast, only 81% conversion was achieved with
Acknowledgment. We thank Stephanie Kramer (UNC-
CH) for her contribution to initial investigations. J.J.B.
thanks the Army Research Office for Staff Research
Funding, and M.R.G. thanks the Department of Energy
(DE-FG02-05ER15630) for their support.
2
þ
2þ
Ru(bpy)3 . While the Ru(dmb)3 catalysts was indeed
faster, the onset of substrate hydrolysis in acetonitrile
reduced the yield, a phenomenon that was exacerbated
1
4ꢀ16
Supporting Information Available. Experimental pro-
cedures and results. This material is available free of
charge via the Internet at http://pubs.acs.org.
with water cosolvent.
promoting ofhydrolysis (CD Cl ), 4 was cleanly converted
Returning to conditions not
2
2
2þ
to 5 without hydrolysis, and Ru(dmb)
17,18
was faster than
3
2
þ
Ru(bpy)3 (Scheme 3).
(
16) These same products have been accessed via metal catalyzed
approaches that similarly proceed through glycosyl radicals. See: Gong,
H.; Andrews, R. S.; Zuccarello, J. L.; Lee, S. J.; Gagn ꢀe , M. R. Org. Lett.
2009, 11, 879–882.
(17) 84% yield for L = dmb (93% brsm); 38% yield for L = bpy
(99% brsm).
(
12) Garrera, H. A.; Gsponer, H. E.; Garcia, N. A.; Cosa, J. J.;
Previtali, C. M. J. Photochem. 1986, 33, 257–260.
13) (a) DeLaive, P. J.; Lee, J. T.; Sprintschnik, H. W.; Abruna, H.;
(
Meyer, T. J.; Whitten, D. G. J. Am. Chem. Soc. 1977, 99, 7094–7097. (b)
DeLaive, P. J.; Foreman, T. K.; Giannotti, C.; Whitten, D. G. J. Am.
Chem. Soc. 1980, 102, 5627–5631.
(18) It is possible the increase in rate with L = dmb is due to an
increased reduction potential (Ered = ꢀ1.47 V compared to ꢀ1.35 V vs
red
=
0
(
14) 75% yield for L = dmb; 54% yield for L = bpy (65% brsm).
15) Full conversion was observed in 6.5 h for both 1 and 5 mol%
SCE). Experiments with L = dtb-bpy (4,4 -di-tert-butyl-bpy, E
(
Ru(dmb)
ꢀ1.68 V) resulted in a 94% conversion and 89% yield (95% brsm) in
2
þ
in 9:1 MeCN/H
3
2
O.
CD Cl over 6 h of irradiation, similar to L = dmb.
2
2
Org. Lett., Vol. 13, No. 9, 2011
2409